A cross-layer design of clustering architecture for wireless sensor networks

Sensor networks enable new applications and pose new challenges on the system designers. In this paper, we investigate sensor networks with emphasis on the design of clustering architecture. The concept of clustering is to partition the numerous sensor nodes into different groups. With clustering, a hierarchical structure can be established in order to allocate resource more efficiently. However, clustering design is a nontrivial task since it involves different protocol layers, including physical layer, scheduling layer and network layer. This paper develops a cluster design method, called "POWER" (power on with elected rotation). The objective of POWER is to determine the optimal number of clusters from the cross-layer aspects of power saving, and coverage performance simultaneously. A closed-from expression for determining the optimal cluster numbers is also derived.